Email Spam Detection

This project presents a comprehensive approach to email spam detection using machine learning algorithms optimized with bio-inspired metaheuristic techniques.With the increasing volume of unsolicited emails, traditional rule based filtering methods have become inadequate, necessitating more sophisticated approaches.Our system implements five machine learning models-Multinomial Na ve Bayes (MNB), Stochastic Gradient Descent (SGD), Decision Tree (DT), Random Forest (RF), and Multi Layer Perceptron (MLP)-enhanced by Particle Swarm Optimization (PSO) and Genetic Algorithm (GA) for parameter tuning.The system was evaluated on seven different email datasets, including the SpamAssassin corpus, Ling-Spam dataset, and Enron Spam dataset.Experimental results show that the MNB classifier optimized with GA achieved the highest accuracy of 100% on the SpamAssassin dataset with an 80:20 train-test split.The integration of bio-inspired algorithms significantly improved classification performance across all models compared to their base implementations.Furthermore, the system includes a user-friendly GUI that allows end users to easily analyze emails and visualize the classification results.This project contributes to the field of email security by demonstrating how bio-inspired optimization techniques can enhance machine learning models' effectiveness in spam detection, potentially reducing the time and resources needed for manual filtering while maintaining high accuracy.